Different Risk Factors and Causes for Early Death after Initiating Dialysis in Diabetic and Non-Diabetic Patients

Abstract

The mortality of patients with end-stage renal disease (ESRD) is especially high after the start of dialysis therapy, especially in diabetic patients. A part of these patients die within three months after initiating renal replacement therapy (RRT). In the present retrospective study we evaluated all patients with ESRD requiring RRT who died within 3 months after initiating the first RRT. A total of 42 patients who died such early after the start of dialysis treatment during the years 1995–2001 were included in the study. Of them, 28 subjects (age 66 + 11 years) were diabetics and 14 non-diabetics (age 76 + 10 years). Indications for the start of dialysis were end-stage renal failure (creatinine clearance < 10–12 mL/min or < 12–14 mL/min in diabetic patients) or fluid lung associated with chronic renal failure (creatinine clearance < 20 mL/min). Hyperhydration with fluid lung was the most common indication for dialysis therapy in patients with diabetes (64.3% versus 14.3%, p < 0.05). The vascular risk factors blood pressure and serum-lipids were similar in both groups; however, diabetic patients were younger than non-diabetic subjects. The prevalence of vascular diseases tended to be higher in the diabetic group, but difference was not significant (see ). Severe heart failure (NYHA stage III-IV) was more common among diabetics (42.8% versus 14.3%, p < 0.05). The incidence of sepsis (17.9% versus 14.3%) did not significantly differ between the groups. The most common cause of death was cardiovascular events in both diabetic and non-diabetic patients (71.5% and 64.2%, respectively). Heart failure was a more common cause of death in diabetic patients (39.2% versus 21.4%, NS). In conclusion, early death after the initiation of dialysis treatment was more common in patients with type 2 diabetes, though, the diabetic patients were less old. In the diabetic group fluid lung was more often indication for initiating dialysis therapy than in the non-diabetic group. In both, diabetic and non-diabetic patients, the most common causes of death are cardiovascular events.

Keywords:

• insulin resistance
• type 2 diabetes
• metabolic control

Introduction

Patients who die within the first three months after commencement of dialysis therapy are not taken into account in health register statistics and may also be excluded from incidence counts.[1] In a recent study,[2] it was shown that at least 6% of all new patients die within 90 days after the initiation of dialysis. Usually the indication for starting dialysis therapy is end-stage renal failure (ESRF) with uremic symptoms and/or a decline in the patient's weekly Kt/V below 2 and GFR below 10–12 mL/min/1.73 m².[3] In patients with chronic heart failure as well as chronic renal insufficiency, the indication for starting dialysis is frequently hyperhydration and/or accumulation of fluid in the lung. Initially, these patients are treated with early dialysis and intensified ultrafiltration or hemofiltration.[4] Independent of the indication for dialysis, diabetic patients on hemodialysis are exposed to a high risk of early death.[5]

The indications for initiating dialysis, the vascular risk factors, the prevalence of vascular diseases, and causes of early death after initiating dialysis therapy were compared in patients with and without diabetes mellitus. The aim of the study was to find out differences in the risk for early death after the start of dialysis between both patient groups.

Figure 1. Prevalence of vascular risk factors (hypertension, hyperlipidemia, and age).

Patients and Methods

A total of 334 patients with ESRF (146 with and 188 without diabetes) started dialysis therapy between 1995 and 2001. Of these, 42 patients died within three months after the start of dialysis treatment. Twenty-eight diabetics and fourteen non-diabetics were included in the study. Thus, early death occurred in 19.1% (28/146) of diabetic and 7.4% (14/188) of non-diabetic individuals requiring chronic dialysis therapy. Indications for the start of dialysis were end-stage renal failure (i.e., creatinine clearance below 10–12 mL/min/1.73 m²; in diabetic patients, 12–14 mL/min/1.73m²) with and without uremic symptoms and/or hyperhydration with fluid in the lung, in conjunction with renal insufficiency (creatinine clearance < 20 mL/min/1.73m2). Baseline characteristics and laboratory parameters of patients with and without diabetes are shown in Table 1.

Table 1 Baseline characteristics of the patients and laboratory data

Patients	Without diabetes (n = 14)	With diabetes (n = 28)
Age (yr)	76 + 10	66 + 11
Sex (F:M)	7:7	12:16
Body mass index	23 + 3	24 + 3
Primary renal diseases (n/%)
Diabetic nephropathy	0	23 (82.1%)
Vascular nephropathy	7 (50%)	3 (10.7%)
Glomerulonephritis	6 (42.8%)	1 (3.6%)
Others	1 (7.1.2%)	1 (3.6%)
Laboratory parameters
S-creatinine (mg/dL)	5.8 + 1.6	5.7 + 1.9
Cr clearance   (mL/min/1.73m2)	10.7 + 2.9	12.6 + 3.4
BUN (mg/dL)	85 + 34	82 + 41
Hemoglobin (g/dL)	10.9 + 1.5	10.1 + 1.1
Serum albumin (g/L)	3.6 + 0.6	3.5 + 0.7
C-reactive peptide (mg/L)	0.7 (0.3–1.4)	0.9 (0.3–1.6)

*p < 0.05.

Primary renal diseases, indications for dialysis, the duration of dialysis therapy, vascular risk factors, the prevalence of vascular diseases, heart failure (NYHA stage II-IV) and the causes of death were compared in patients with and without diabetes. The frequency of late referral to the nephrology department was also evaluated.

The diagnosis of diabetic nephropathy was based on the onset of persistent macro-proteinuria with normal urine sediment, normal kidney sonogram, long-acting diabetes, and concomitant diabetic retinopathy. Glomerulonephritis was established by renal biopsy. Vascular nephropathy stands for a heterogenous group of patients with vascular diseases and is also referred to as ischemic nephropathy.[6],[7] In the present study, the diagnosis of vascular disease was based on the demonstration of a normal urine sediment without proteinuria and reduced kidney size in the ultrasonography investigation. In two cases, the diagnosis was additionally established by renal biopsy. Heart failure was diagnosed on the basis of clinical data and fluid congestion on the lung x-ray.

Hyperhydration was defined as fluid overload of the body, usually associated with only slightly impaired cardial function in echocardiography. Coronary artery disease was diagnosed by the demonstration of ischemic changes on electrocardiogram (ECG) and/or myocardial infarction in the patient's medical history. Peripheral vascular disease was established when the patient had a history of claudication and/or amputation. The causes of death were evaluated on the basis of autopsy in 36 patients and the information provided by the general physician in six cases.

Statistical Analysis

Student's t-test was used to compare differences between groups and the chi-square test to compare prevalences. The level of statistical significances was set at p < 0.05.

Results

Hyperhydration and the accumulation of fluid in the lung was the most common indication for dialysis therapy in patients with diabetes (64.3% versus 14.3%, p < 0.05). In patients without diabetes, the most common primary renal disease was vascular nephropathy (50%). The prevalence of hypertension and hyperlipidemia as vascular risk factors was similar in both groups; however, diabetic patients received statin therapy more frequently (64.2% versus 28.5%, p < 0.05). On the other hand, the diabetics were younger than the non-diabetic subjects (76 ± 10 versus 66 ± 11 years). The prevalence of heart failure (53.5% versus 42.8%) and that of all vascular diseases were higher among diabetic patients, but differences were not significant. However, severe heart failure (NYHA stage III-IV) was more common among diabetics (42.8% versus 14.3%, p < 0.05). The prevalence of all risk factors is summarized in Table 2. The number of dialysis sessions per patient was approximately the same among those with (11 ± 6) and without (12 ± 7) diabetes. The incidence of sepsis (17.9% versus 14.3%) as an additional prognostic factor did not significantly differ between the groups.

Table 2 Vascular risk factors, prevalence of heart failure, and vascular disease in patients with and without diabetes

Patients	Without diabetes	With diabetes
Vascular risk (x + SD)
Fasting blood glucose (mg/dL)	92 + 12^†	132 + 19
HbA1c (%)	5.6 + 0.3^†	7.4 + 1.3**
Cholesterol (mg/dL)	202 + 34	218 + 46
Triglycerides (mg/dL)	166 + 56	178 + 65
Systolic blood pressure (mm Hg)	135 + 16	136 + 22
Diastolic blood pressure (mm Hg)	86 + 9	67 + 9
Prevalence of vascular risk factors (n/%)
Hypertension	11 (78.5%)	23 (82.1%)
Antihypertensive drugs   (n + range)	3 (0–4)	3 (0–4)
ACE inhibitors	7 (50%)	21 (75%)
Hypercholesterolemia	7 (50%)	23 (82,1%)
Statine therapy	4 (28.5%)	18 (64.2%)
Heart failure (n/%)
NYHA stage II-IV	6 (42.8%)	15 (53.5%)
NYHA stage III-IV	2 (14.3%)	12 (42.8%)
Vascular diseases (n/%)
Coronary heart disease	6 (42.8%)	16 (53.5%)
Peripheral vascular disease	2 (14.3%)	11 (39.3%)

*p < 0.05, ^†p < 0.01.

The most common cause of death was cardiovascular events in both diabetic and non-diabetic patients (71.5% and 64.2%, respectively). Heart failure was a more common cause of death in diabetic patients (39.2% versus 21.4%, NS). The prevalence of all causes of death in the two groups is summarized in Table 3. Late referral to the nephrology department (referral fewer than six months before the start of dialysis) was frequently in both the diabetic patient group (75%) and the non-diabetic group (57.1%). In the diabetic patients, starting dialysis happened more frequently in the dialysis unit (71.4%) than in the intensive care unit. The first treatment session was performed via central venous access in most cases (88.1%).

Table 3 Indications for dialysis in the patients with and without diabetes

Patients	Without diabetes	With diabetes
Indications for dialysis (n/%)
Hyperhydration with fluid lung	2 (14.3%)	18 (64.3%)
Uremic symptoms	6 (42.8%)	4 (14.2%)
Catabolism	4 (28.6%)	2 (7.1%)
Planned (Cr clearance < 12 mL/min)	2 (14.3%)	0
Causes of death (n/%)
Heart failure	3 (21.4%)	11 (39.2%)
Myocardial infarction	3 (21.4%)	7 (25%)
Sepsis	2 (14.3%)	5 (17.9%)
Pneumonia	3 (21.4%)	2 (7.1%)
Sudden death	3 (21.4%)	2 (7.1%)
Stroke	0	1 (3.6%)

*p < 0.05.

Discussion

Mortality among ESRF patients undergoing renal replacement therapy (RRT) remains high, especially in those with type 2 diabetes.[6] A relatively large number of these patients die shortly early after the initiation of RRT. High mortality within the first three months after the start of dialysis therapy is more common in diabetic patients. Usually individuals who die early under dialysis are registered as acute, and not chronic, ESRF patients. Therefore, these patients are not included in most national health register statistics in Europe.[5] Similarly, in the United States, patients who die within 90 days after the start of dialysis are not included in the mortality statistics of health registers.[2] In the present study, early death occurred in 12%, which is much higher than 6% as reported in the literature. This higher mortality in this study may be explained by the fact that all patients were included who started dialysis therapy due to pulmonary edema or other complications and who died within 90 days after initiation of dialysis. It cannot be excluded that in some patients, a partial restoration of renal function would have been developed but patients died before restoration.

The majority of the patients who died early after the start of dialysis therapy were diabetic patients (19.1% versus 7.4% non-diabetics). The high incidence of early death in the diabetic patients may be explained by the fact that all patients who were dialyzed due to fluid lung at the presence of advanced renal insufficiency were evaluated in this study. In these patients, creatinine clearance was usually < 20 mL/min/1.73m2; however, one of the diabetic patients with fluid lung needed intermittent dialysis therapy at a clearance of 26 mL/min/1.73m². Hyperhydration with fluid in the lung was more frequently the indication for early dialysis therapy in patients with diabetes. The mean age of diabetic patients was lower than that of non-diabetic individuals.

Predictors for early death in patients with end-stage renal disease (ESRD) have been evaluated in several studies. In a recent report, nutritional indices prior to the initiation of RRT and a previous history of cardiovascular disease were the best predictors of early death.[8] Generally, vascular complications also were an important cause of early death in ESRD patients,[1] particularly patients on dialysis who have severe coronary heart disease and acute myocardial infarction die due to cardiac causes.[9] Additional important predictors of early mortality in patients on hemodialysis are low serum albumin levels and reduced BMI[10] as well as elevated C-reactive peptide (CRP) levels.[11] In patients on peritoneal dialysis, rapid loss of residual renal function is an additional important predictor of early mortality.[12] In the current diabetic group, five patients had diabetes but not diabetic nephropathy. To mix dialysis patients with concomitant diabetes and patients with diabetic nephropathy is suitable because the prognosis is similarly poor in both patient groups.[13]

Coronary heart disease is the most common cause of morbidity in ESRD.[4] Usually the prevalence of traditional coronary risk factors is high in uremic patients. However, the extent of mortality is so high that non-traditional risk factors may also contribute to coronary heart disease in ESRF.[14] During the last few years, several studies have reported the acute effects of hemodialysis on endothelial function and the elasticity of large arteries. These effects may contribute to vascular disease in ESRF patients.[15],[16] In this study, the prevalence of hypertension and hyperlipidemia as vascular risk factors for cardiovascular mortality was similarly high in patients with and without diabetes; however, the diabetic subjects received more frequent statin therapy. Therefore, no significant difference in lipid status could be observed. While the prevalence of heart failure, including all stages, and coronary heart disease tended to be higher in patients with diabetes, the prevalence of severe heart failure NYHA III-IV was significantly higher in the diabetic group. BMI and residual renal function before the start of dialysis were higher in the diabetic patients; however, the differences were not statistically significant

In this study, the majority of the patients who died early after the start of dialysis were diabetics, and hyperhydration with a fluid lung was the most common indication for early dialysis in patients with diabetes. The high number of our diabetic patients with early death can be explained by the including of patients with fluid lung as indication for dialysis therapy. In both diabetics and non-diabetics, the most common causes of death were cardiovascular events. Heart failure was more frequently a cause of death in diabetic patients; however, the difference was not statistically significant.

Much of the literature demonstrates the disadvantage of “late referral” and the advantage of care by nephrologists with expertise in the multi-morbidity of patients with renal diseases,[17] though it is important to cooperate with other health professionals.[18] However, it is the nephrologist's job to integrate all the findings to the one comprehensive treatment, including ultrafiltration via a dialysis machine, which is one of the most efficient techniques to treat overhydration or pulmonary edema in decompensated heart failure with only slightly impaired renal function.[19]

In conclusion, early death after the initiation of dialysis treatment was more common in patients with type 2 diabetes. At the start of dialysis, the diabetic patients were younger than non-diabetic subjects. In the diabetic group, hyperhydration with fluid in the lung due to severe myocardial failure was more frequently the indication for initiating dialysis therapy than in the non-diabetic group. When comparing both groups, blood glucose was higher in the diabetics and age was higher in the non-diabetic group. Cholesterol levels were similar in both groups; however, diabetic patients needed more frequently statin therapy. In both diabetic and non-diabetic patients, the most common causes of death were cardiovascular events. The high number of early death in the diabetic group implicates better management of these patients before starting renal replacement therapy.